Study Methods Tip: A Good Sample Defines Your Study

By Scott D. Crawford, Director of Data Collection & Founder of SoundRocket

There is no better example of the old axiom—you get out what you put it—than when applied to the selection of a sample for your survey data collection. Selecting the right sample will ensure that you can answer the questions that you desire when your study is done.

Sampling can be a scary process for those who are not familiar with the techniques, so when possible, we do recommend connecting with a professional sampling statistician to assist. However, there are some basics to consider when implementing a study like the MSL.

At the core of most sampling is the concept of a simple random sample. If you have 10,000 students enrolled at your University, and you would like to select a simple random sample of 4,000 students, that means that you select students from the overall population randomly (using a random function) until you have selected 4,000 total. The idea is that in surveying these students, you will have data that will represent the full population of 10,000 students enrolled.

The MSL is designed on the basic understanding that each school will select a random sample of students, providing the ability to compare students between institutions (or groups of institutions). The result is that you can say “Students enrolled at our University scored higher on resiliency than students at other schools like ours.” If your sample was not random (or other school samples were not random), you would not be able to do this comparison as easily (if at all).

The simple random sample is not the only path for MSL participation. So long as a school can maintain the core purpose of the study with a simple random sample, options for oversampling also exist. In MSL terms, we call these Supplemental Samples, and they may vary considerably. Some options for Supplemental Samples used by schools in the past include:

• Oversampling of Program Participants. If your institution has a leadership program that you would like to have included in the study, these individuals may be included. Once your random sample is selected, you can create a supplemental sample that brings in the remaining leadership program participants. This way you can make comparisons between your leadership students and other students at your institutions, or even students at other schools.

• Oversampling of Specific Demographic Groups. If you wish to have a larger sample of an underrepresented minority at your institution, you may also add a supplemental sample to do so. In the MSL, we can assist you in determining the best strategy to oversample, whether you are taking a larger portion of, or even a full population of, the target demographic category. Many schools have done this to allow for more complete data on specific populations.

There is a lot of room to build on a base simple random sample to enhance the analytic power of your study! Talk with a sampling statistician, or if you are participating in the MSL, as your study coordinator for assistance if you would like to pursue these options. For most enhanced sampling in the MSL, there is no added cost.